INDEX

Explanations

phrases indicating types and categories

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

findpost

-0.49

 CURLOPT

-0.47

 cérami

-0.47

 oration

-0.44

ropractor

-0.44

 diner

-0.43

 Ceremony

-0.43

愷

-0.42

 vœux

-0.42

 AppComponent

-0.42

POSITIVE LOGITS

 populations

0.62

 kinds

0.56

 leerlingen

0.53

 types

0.52

Lugares

0.50

 members

0.49

 leaders

0.48

 adherents

0.47

 servants

0.47

 themselves

0.47

Activations Density 0.364%

phrases indicating types and categories

No Comments

No Known Activations